Iceberg Ahead: dbt now supports Apache Iceberg tables in BigQuery

We’ve got cool news—literally.

With the latest release of the dbt-bigquery adapter, you can now materialize Iceberg tables directly in BigQuery. That’s right: your dbt models can now land in an open table format designed for massive scale, multi-engine interoperability, and metadata-rich lakehouse architectures.

This unlocks new doors for organizations that embrace open formats, AI governance, and cross-platform compatibility—all while staying within the comfort of dbt.

Let’s unpack what this means.

First, what even is a catalog?

These days, everyone in data is talking about catalogs. But depending on who you ask, a “catalog” might be a UI, a compliance tool, a metadata layer, or a magic box that solves data governance.

So let’s keep it simple.

A data catalog is a centralized metadata layer that helps humans and machines understand what data exists, where it lives, and how to interact with it.

In dbt’s world, that means knowing what assets already exist, where to materialize new ones, and how different engines (like BigQuery or Spark) can discover those models.

What makes Iceberg special?

Apache Iceberg is an open table format purpose-built for analytics at scale. Think of it as a smarter way to manage massive datasets—complete with time travel, schema evolution, and lightning-fast metadata operations.

An Iceberg catalog serves as the logical registry of your tables. It tells compute engines how to find and interact with the underlying files. In BigQuery, that catalog is now fully supported via dbt.

Google provides documentation around their implementation here.

Why should dbt users care?

For dbt developers, supporting Iceberg in BigQuery isn’t just a nice-to-have—it’s a powerful unlock. Here’s why:

• Discover faster: With better metadata awareness dbt can register models in a catalog so that other tools (and dbt itself) can discover them.
• No extra copies: Cross-engine compatibility so a table built in BigQuery via dbt can now be queried by Spark, Trino, or other Iceberg-aware engines.
• Greater flexibility: Dynamic execution means that dbt compiles code based on what already exists. Iceberg catalog support makes that compilation smarter and more flexible in modern lakehouse environments.

How it works in dbt

Using the new Iceberg support is simple and consistent with other dbt workflows:

1. Create a catalogs.yml file in the top level of your dbt project:

2. You’ll then declare Iceberg-specific configs like this in your model:

3. Execute the dbt model with a dbt run -s iceberg_model

The full starter guide can be found here.

That’s it. Behind the scenes, the dbt-bigquery adapter will:

• Create the table as an Iceberg table in BigQuery
• Register it in the configured catalog
• Allow you to manage and query it just like any other dbt model

🔐 Note: You’ll need the BigQuery Connection Admin role to make this work, as the adapter must register the table with metadata.

What's next?

This is just the beginning. As dbt’s semantic layer, Mesh, and multi-platform story continue to evolve, open formats like Iceberg will play a significant role in building AI-ready data systems. We're actively exploring:

• Auto-generating storage_uri from your project + model structure
• Configurable default catalogs per environment
• More granular access control for metadata management

dbt + BigQuery + Iceberg = A lakehouse dream team

This update represents a significant step forward for modern data teams, offering flexible open formats, native BigQuery support, and a dbt-native approach to manage it all.

Whether you're building data products, governed metrics, or AI-ready pipelines, Iceberg support in dbt + BigQuery brings speed, scale, and structure to your lakehouse strategy.

Ready to try it? Update to the latest version of dbt, claim your permissions, and let the dbt run begin.